I’ve borrowed my title from Marcie Wheeler’s book [Anatomy of Deceit]. She wrote a detailed narrative of the Bush Administration’s lying us into the Invasion of Iraq and the cover-up that followed, specifically involving Joseph Wilson and his wife, Valerie Plame – the outed CIA Agent. Marcie Wheeler is a Comparative Literature Ph.D. who wrote her doctoral thesis on what amounted to ancient blogs – short newspaper essays. During the Bush years, she kept us all informed, blogging as emptywheel on FireDogLake. She’s still at it here. She is a master of close-reading of texts and using case study to dissect much larger issues. While her writing was not widely known, she was the ‘source’ for many of the things in the Mainline Press that lead the country to our current general understanding of what went on in those years.
Suicidal Thoughts and Behavior With Antidepressant Treatment
Reanalysis of the Randomized Placebo-Controlled Studies of Fluoxetine and Venlafaxine
by Robert D. Gibbons, PhD; C. Hendricks Brown, PhD; Kwan Hur, PhD; John M. Davis, MD; and J. John Mann, MD
Archives of General Psychiatry. Published online February 6, 2012.
…Conclusions: Fluoxetine and venlafaxine decreased suicidal thoughts and behavior for adult and geriatric patients. This protective effect is mediated by decreases in depressive symptoms with treatment. For youths, no significant effects of treatment on suicidal thoughts and behavior were found, although depression responded to treatment. No evidence of increased suicide risk was observed in youths receiving active medication. To our knowledge, this is the first research synthesis of suicidal thoughts and behavior in depressed patients treated with antidepressants that examined the mediating role of depressive symptoms using complete longitudinal person level data from a large set of published and unpublished studies.
Dr. Gibbons, who was on the FDA’s advisory committee that voted in favor of the black box warning, said he was very concerned about the validity of the data that prompted the affirmative vote. "The adverse event reports for suicidal thoughts and behavior showed a fairly small signal in children, but the prospective measurements showed no effect of treatment whatsoever. As a statistician, I put more weight in prospective data than these retrospective reports," he said. "The vote was 15 to 8, so some of the members of the committee, myself included, were not persuaded by those reports and felt that the black box warning was not warranted"…
But I can’t really get away with accusing Gibbons with being obsessed with the topic because I did the same thing over the month that followed – becoming obsessed myself with the Gibbons’ meta-analysis:
| Feb 09 | a book review… | Feb 22 | significant II… | Feb 25 | watchful waiting… |
| Feb 10 | sometimes… | Feb 23 | significant III… | Feb 29 | smell a campaign… |
| Feb 22 | significant I… | Feb 24 | coming soon?… | Mar 04 | red alert… |
Here’s what kept me at it besides a suspicion of bias based on his history. The article had no data or specific reference to the data. I was only interested in the child studies. He said the data was from the complete data from the NIMH TADS study, all the Lilly Trials of Prozac and all the data from the adult Wyeth trials of Effexor and that the details would be in a companion article coming later. I knew that there were two Wyeth trials of Effexor in kids. Why weren’t they there? He said there were 4 child trials in his meta-analysis. I could only account for three – TADS and the two Lilly Trials submitted to the FDA [X065 and HCJE]. The statistical analysis itself was described in detail, but was leagues above the heads of the audience [psychiatrists, me included]. I wrote asking for the data sources and his reply was coming-in-the-next-article. No data? vague references for the data source? unintelligible analysis? terse results? A topic already meta-analyzed ad nauseum with different results? And on top of that, there were interviews and commentaries on his results on NPR, Medscape, in the newspapers – and they were pretty explicit:
“hidden in the fog”…
seems like there’s been a lot of that going around lately,
duane
Dr. Gibbons was quoted as follows: “I hope that the warnings will not prevent depressed children and adults from getting treatment for depression… The greatest cause of suicide is untreated or undiagnosed depression. It’s very important that this condition be recognized and appropriately treated and not discarded because doctors are afraid to be sued.â€
We could spend a semester deconstructing the logical weaknesses of this statement. For starters, let me just point out that undiagnosed depression is a red herring in the present context, as the decision to prescribe or to not prescribe the drugs and to warn about their potential side effects is predicated on an existing diagnosis.
Second, Dr. Gibbons should know – if he is going to make such statements – that there is no good evidence that antidepressant drugs do prevent suicide. The WHO looked into this question in 2004 and came up empty except for some evidence that lithium has that property: http://www.euro.who.int/__data/assets/pdf_file/0010/74692/E83583.pdf
The most vocal proponent of the claim that antidepressant drugs prevent suicide has been a Swedish researcher named Goran Isacsson. His most recent publication, which already was marked by dubious logic, had to be retracted: http://tinyurl.com/7er7h28
And then there is the issue of suicide attempts in the TADS study of adolescents. As Dr. Nardo pointed out here http://1boringoldman.com/index.php/2012/02/22/significant-i/ with a clarifying graphic, 17 of 18 attempts in this study occurred in children who were taking fluoxetine (Prozac) at the time. That is hardly a ringing endorsement of Gibbons’ position.
Come to think on it, what are Gibbons’ credentials for pontificating on FDA policy or clinical guidelines? Is he a clinician? No. Is he trained in clinical psychopharmacology? No. Does he make clinical decisions as a professional statistician? No. Does he appear to have an agenda? Yes. Is he grandstanding? Yes. I will refrain from speculating on the dynamics of his bias and his agenda, and I will just end by saying Dr. Gibbons should stick to his statistical knitting and keep out of issues on which he is unqualified to pontificate.
i’m immensely grateful to people with the requisite knowledge who hunt down and expose bad science for the benefit of the rest of us!
Dr. Carroll, I’ve always wondered about “undiagnosed depression” as a cause of suicide. If it’s undiagnosed, it would have to be diagnosed post-mortem, correct? Perhaps from hearsay?
Very easy to project one’s own construction on the emotional state of a dead person.
Alto: there is a modest literature on ‘psychological autopsy’ in the study of ‘undiagnosed depression.’ There are methodologic expectations, and I think there is some validity to this approach but I would not hold it out as absolutely reliable. After all, even in the living there is palpable unreliability of diagnosis (coefficients of contemporaneous diagnostic concordance around 0.75 for major depression, and estimates of diagnostic stability over time are even lower).